INDEX

Explanations

scientific or factual evidence

oai_token-act-pair · gpt-3.5-turbo

references to evidence and the strength of such claims

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

•Evidence

No Comments

Negative Logits

ategory

-0.73

ttle

-0.73

otom

-0.68

Hop

-0.67

ernel

-0.66

skill

-0.65

 throats

-0.64

jug

-0.64

awar

-0.63

quer

-0.62

POSITIVE LOGITS

evidence

1.07

 evidence

1.03

Evidence

0.95

 Evidence

0.92

 conclusive

0.80

 corrobor

0.78

edly

0.78

 evid

0.77

proof

0.77

 suggests

0.76

Activations Density 0.031%

scientific or factual evidence

references to evidence and the strength of such claims

No Comments

No Known Activations