INDEX

Explanations

words related to things that are difficult to understand, or not well known or prominent

oai_token-act-pair · gpt-3.5-turbo

References to obscure or arcane knowledge and concepts

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

Demand

-0.79

iphate

-0.76

©¶æ

-0.71

MIC

-0.70

avers

-0.70

hao

-0.69

ynthesis

-0.68

uana

-0.67

Adam

-0.66

onz

-0.66

POSITIVE LOGITS

 obscure

0.95

ly

0.87

 trivia

0.80

uously

0.76

ively

0.75

ulously

0.73

worldly

0.72

ingly

0.72

 footnote

0.72

 escapes

0.70

Activations Density 0.019%

words related to things that are difficult to understand, or not well known or prominent

References to obscure or arcane knowledge and concepts

No Comments

No Known Activations