INDEX

Explanations

adjectives and nouns related to conflict or hostility

oai_token-act-pair · gpt-3.5-turbo

words and suffixes indicating negative or hostile traits and actions

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

herty

-0.76

onom

-0.73

frey

-0.73

mist

-0.72

mans

-0.72

icter

-0.71

ilts

-0.70

sav

-0.69

ð

-0.66

orah

-0.66

POSITIVE LOGITS

 thereto

0.93

 Pact

0.68

 Covenant

0.68

train

0.66

 Royale

0.64

 pact

0.64

Ake

0.63

 Territories

0.63

backer

0.62

 Agreement

0.61

Activations Density 0.185%

adjectives and nouns related to conflict or hostility

words and suffixes indicating negative or hostile traits and actions

No Comments

No Known Activations