INDEX

Explanations

phrases related to sending a message or signaling something

oai_token-act-pair · gpt-3.5-turbo

statements related to proving capability or credibility

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Medline

-0.69

 Torrent

-0.61

blogspot

-0.60

anchester

-0.58

rouse

-0.58

 PubMed

-0.58

 clust

-0.57

 cardinal

-0.57

idious

-0.57

 whirlwind

-0.56

POSITIVE LOGITS

Hey

0.83

worthiness

0.82

chwitz

0.81

 superiority

0.79

iors

0.73

 inferior

0.73

hey

0.72

 conformity

0.71

 Compliance

0.69

enance

0.67

Activations Density 0.525%

phrases related to sending a message or signaling something

statements related to proving capability or credibility

No Comments

No Known Activations