INDEX

Explanations

proper nouns and names and references to specific media

oai_token-act-pair · gpt-3.5-turbo

statements or phrases indicating definitive actions or conclusions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Jace

-0.93

 ponies

-0.91

 glyph

-0.91

 phil

-0.89

 crypt

-0.89

Peg

-0.89

gem

-0.88

gem

-0.87

 Crypt

-0.87

 whip

-0.84

POSITIVE LOGITS

 Anderson

2.75

Anderson

2.65

 Andersen

1.23

 Weaver

1.07

ERSON

1.03

 Omaha

0.99

 Amanda

0.96

 Silva

0.96

 Arnold

0.94

 Weed

0.91

Activations Density 0.214%

proper nouns and names and references to specific media

statements or phrases indicating definitive actions or conclusions

No Comments

No Known Activations