INDEX

Explanations

verbs associated with refusal or denial

oai_token-act-pair · gpt-3.5-turbo

instances of the contraction "won't"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Kings

-0.68

culated

-0.68

itiz

-0.64

 illustration

-0.63

soType

-0.62

 Floating

-0.59

FontSize

-0.58

 Anat

-0.58

 Metallic

-0.58

 contrasting

-0.58

POSITIVE LOGITS

 necessarily

0.98

ember

0.84

rue

0.82

entimes

0.78

bud

0.78

erest

0.77

payers

0.77

acters

0.75

angular

0.75

ardless

0.74

Activations Density 0.036%

verbs associated with refusal or denial

instances of the contraction "won't"

No Comments

No Known Activations