INDEX

Explanations

phrases related to official statements or announcements

oai_token-act-pair · gpt-3.5-turbo

references to press and news conferences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

hetti

-0.64

 Honest

-0.63

 Monk

-0.60

 Marble

-0.60

 perfected

-0.58

 Ness

-0.58

 Friendship

-0.58

 Tavern

-0.57

 Butcher

-0.57

 Judd

-0.57

POSITIVE LOGITS

prise

0.88

 conference

0.83

conference

0.83

 query

0.81

icer

0.75

 release

0.74

agency

0.73

 statement

0.71

release

0.69

 briefing

0.69

Activations Density 0.047%

phrases related to official statements or announcements

references to press and news conferences

No Comments

No Known Activations