INDEX

Explanations

negations or refusals

oai_token-act-pair · gpt-3.5-turbo

negations or terms that imply a lack or absence

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

 Carbuncle

-0.65

assador

-0.64

Dear

-0.63

asma

-0.61

 nonetheless

-0.61

MpServer

-0.60

è¦ļéĨĴ

-0.59

arthy

-0.58

CI

-0.58

 effectiveness

-0.56

POSITIVE LOGITS

 hesitate

1.11

shy

0.94

 disappoint

0.90

 forgetting

0.81

 overlooked

0.81

 hesitated

0.76

 doubt

0.76

 afraid

0.75

 forget

0.74

 precon

0.72

Activations Density 0.822%

negations or refusals

negations or terms that imply a lack or absence

No Comments

No Known Activations