INDEX

Explanations

simple tasks or concepts

oai_token-act-pair · gpt-3.5-turbo

references to simplicity or simplification in various contexts

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

psey

-1.12

etheus

-0.72

outheast

-0.72

 whale

-0.71

 Normandy

-0.69

 Whale

-0.68

reon

-0.66

everal

-0.66

ificantly

-0.65

 whales

-0.64

POSITIVE LOGITS

 arithmetic

0.84

ItemImage

0.74

tons

0.73

simple

0.69

 simple

0.67

execute

0.67

tru

0.65

PLAY

0.65

ASY

0.64

 elegance

0.64

Activations Density 0.275%

simple tasks or concepts

references to simplicity or simplification in various contexts

No Comments

No Known Activations