INDEX

Explanations

phrases or sentences asking for information or decisions

oai_token-act-pair · gpt-3.5-turbo

questions or phrases beginning with "what."

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

ulic

-0.75

enburg

-0.71

ubs

-0.69

robe

-0.69

ped

-0.67

ster

-0.67

enberg

-0.65

eah

-0.64

aches

-0.63

trop

-0.63

POSITIVE LOGITS

 happened

1.23

 happens

1.14

soever

1.14

 kinds

1.07

 sorts

1.07

 happ

1.06

 transpired

1.04

 else

0.93

 exactly

0.90

 constitutes

0.88

Activations Density 0.104%

phrases or sentences asking for information or decisions

questions or phrases beginning with "what."

No Comments

No Known Activations