INDEX

Explanations

phrases related to topics like phenomena, PC sales, details, predecessors, climate change, human-induced climate change, detainees, marriage definitions, high-ranking officials, modern America, definitions, US officers, government programs, advances, robotic surgery, whispering, apples, winners, finalists, forces, and

oai_token-act-pair · gpt-3.5-turbo

terms related to phenomena and definitions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bilt

-0.63

srf

-0.58

 Leilan

-0.58

ahime

-0.56

inis

-0.53

ipel

-0.53

allery

-0.53

quished

-0.52

;;;;

-0.52

ramid

-0.52

POSITIVE LOGITS

 should

0.60

 might

0.60

 cannot

0.60

 hadn

0.59

 exists

0.58

 could

0.57

 existed

0.56

)]

0.56

 couldn

0.56

 lacks

0.56

Activations Density 1.414%

terms related to phenomena and definitions

No Comments

No Known Activations