INDEX

Explanations

news-related terms

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "news."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ignt

-0.71

 sshd

-0.66

 Phill

-0.64

ukong

-0.64

 Cheong

-0.64

 grun

-0.63

aughs

-0.62

 Hamm

-0.61

 Strauss

-0.60

 Philippe

-0.60

POSITIVE LOGITS

flash

1.06

room

0.97

rooms

0.91

letters

0.91

reader

0.85

worthy

0.84

LET

0.81

worthiness

0.81

ource

0.80

agents

0.79

Activations Density 0.027%

news-related terms

repeated mentions of the word "news."

No Comments

No Known Activations