INDEX

Explanations

individuals providing information or quotes in statements or interviews

oai_token-act-pair · gpt-3.5-turbo

the word "told" indicating reporting or communication of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 ILCS

-0.86

icons

-0.71

reci

-0.66

abilities

-0.64

estern

-0.64

spir

-0.64

adesh

-0.64

engeance

-0.63

oppable

-0.63

otion

-0.63

POSITIVE LOGITS

tale

1.03

tell

0.84

ingly

0.82

me

0.79

him

0.76

 bluntly

0.75

warn

0.75

us

0.74

told

0.72

 reporters

0.72

Activations Density 0.057%

individuals providing information or quotes in statements or interviews

the word "told" indicating reporting or communication of information

No Comments

No Known Activations