INDEX

Explanations

sentences containing reported speech where someone is talking or relaying information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "said" indicating dialogue or conversations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ptoms

-0.83

wx

-0.82

=~=~

-0.76

irtual

-0.74

pering

-0.73

xtap

-0.73

pleting

-0.72

isoft

-0.71

cv

-0.71

avez

-0.70

POSITIVE LOGITS

 goodbye

1.47

 hello

1.36

yes

1.20

bye

1.18

 nothing

1.11

hey

1.09

hi

1.06

 yeah

1.03

 aloud

1.01

 something

0.98

Activations Density 0.136%

sentences containing reported speech where someone is talking or relaying information

instances of the word "said" indicating dialogue or conversations

No Comments

No Known Activations