INDEX

Explanations

phrases related to halting or stopping actions or processes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "suspend" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Kaiser

-0.73

ographer

-0.65

 Empire

-0.65

gal

-0.64

omen

-0.64

ortion

-0.63

CTR

-0.63

 Whitney

-0.62

NYC

-0.62

obi

-0.61

POSITIVE LOGITS

 suspend

3.68

 suspending

2.61

susp

2.42

 susp

2.13

 Susp

2.07

usp

1.90

 suspension

1.88

 suspensions

1.85

 suspended

1.81

 revoke

1.17

Activations Density 0.013%

phrases related to halting or stopping actions or processes

instances of the word "suspend" and its variations

No Comments

No Known Activations