INDEX

Explanations

words related to statements, promises, or actions made by politicians or public figures

oai_token-act-pair · gpt-3.5-turbo

the use of a specific character or symbol commonly found in text

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

 layered

-0.76

 consequential

-0.74

SAM

-0.72

 territory

-0.66

 Rica

-0.66

 moments

-0.65

 stretched

-0.65

 wrinkles

-0.64

 Chandler

-0.63

 Maher

-0.63

POSITIVE LOGITS

replace

1.08

create

1.08

educ

1.07

raise

1.05

remove

1.05

delete

1.04

reci

1.03

lean

1.03

activate

1.02

spons

1.02

Activations Density 0.069%

words related to statements, promises, or actions made by politicians or public figures

the use of a specific character or symbol commonly found in text

No Comments

No Known Activations