INDEX

Explanations

contractions of words

oai_token-act-pair · gpt-3.5-turbo

negative contractions and expressions of uncertainty or lack of knowledge

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

å£«

-0.74

Topics

-0.68

 McDonnell

-0.68

Nine

-0.66

Pope

-0.65

 heavens

-0.65

ãĤ£

-0.65

 Sudan

-0.63

Gaza

-0.62

Reviewer

-0.60

POSITIVE LOGITS

 dont

0.99

 wont

0.81

 didnt

0.80

rontal

0.77

hes

0.76

cha

0.75

 alot

0.73

 dunno

0.70

creen

0.68

 blackout

0.67

Activations Density 0.025%

contractions of words

negative contractions and expressions of uncertainty or lack of knowledge

No Comments

No Known Activations