INDEX

Explanations

continuations to news stories or articles

oai_token-act-pair · gpt-3.5-turbo

phrases indicating ongoing narratives or segments in text

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

Not in Any Lists

No Comments

Negative Logits

aster

-0.67

wives

-0.66

 majesty

-0.66

ĪĴ

-0.66

onomic

-0.65

asters

-0.65

ofi

-0.64

 charismatic

-0.64

ogy

-0.64

uras

-0.63

POSITIVE LOGITS

 Continued

1.06

 Transcript

0.91

 VIDEOS

0.89

 Below

0.79

 ..........

0.74

...]

0.73

CLASSIFIED

0.72

},"

0.72

 Emails

0.71

 Loading

0.67

Activations Density 0.006%

continuations to news stories or articles

phrases indicating ongoing narratives or segments in text

No Comments

No Known Activations