INDEX

Explanations

specific actions or events in a narrative context, often related to news or storytelling

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions or events that contribute to a narrative development

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

omever

-0.46

See

-0.44

enegger

-0.44

Make

-0.44

Template

-0.43

GROUP

-0.42

 cohorts

-0.41

Individual

-0.41

Unless

-0.40

Non

-0.40

POSITIVE LOGITS

 yesterday

0.68

nesday

0.64

 aback

0.56

emonium

0.53

 applause

0.52

 landfall

0.51

wolves

0.49

ĸļ

0.49

 Â£

0.48

yss

0.48

Activations Density 0.581%

specific actions or events in a narrative context, often related to news or storytelling

phrases indicating actions or events that contribute to a narrative development

No Comments

No Known Activations