INDEX

Explanations

passages indicating that something has been written

oai_token-act-pair · gpt-3.5-turbo

instances of the word "written" in various contexts

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

abe

-0.86

Ĭ±

-0.85

nel

-0.82

ugal

-0.80

 Shinra

-0.75

elli

-0.73

illon

-0.73

allows

-0.72

isters

-0.71

alin

-0.70

POSITIVE LOGITS

 essays

0.75

 instrument

0.73

 breath

0.72

acters

0.71

 essay

0.70

 excerpts

0.69

 written

0.69

 typed

0.68

 aloud

0.67

itatively

0.66

Activations Density 0.025%

passages indicating that something has been written

instances of the word "written" in various contexts

No Comments

No Known Activations