INDEX

Explanations

references to the physical throat or actions related to it

oai_token-act-pair · gpt-3.5-turbo

references to the throat and related contexts, often in a negative or violent manner

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

udeb

-0.66

 Scient

-0.65

notation

-0.63

AUT

-0.63

ARC

-0.58

vol

-0.57

occ

-0.57

atern

-0.57

 prudent

-0.56

amazon

-0.56

POSITIVE LOGITS

 throat

1.46

 throats

1.32

pipe

0.96

 cavity

0.91

 slit

0.91

ħĭ

0.89

bone

0.85

piece

0.83

thro

0.83

odox

0.81

Activations Density 0.006%

references to the physical throat or actions related to it

references to the throat and related contexts, often in a negative or violent manner

No Comments

No Known Activations