INDEX

Explanations

words related to significant effects or consequences

oai_token-act-pair · gpt-3.5-turbo

mentions of various impacts on different aspects of life or the environment

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

 confess

-0.78

unny

-0.72

 confession

-0.65

 arranged

-0.62

mah

-0.62

 Playboy

-0.62

tumblr

-0.61

strate

-0.60

 Style

-0.60

 secretly

-0.59

POSITIVE LOGITS

 impacts

3.57

 impact

2.34

impact

2.08

 impacting

2.00

 impacted

1.90

 affects

1.86

 Impact

1.80

 effects

1.74

 harms

1.64

 ramifications

1.60

Activations Density 0.015%

words related to significant effects or consequences

mentions of various impacts on different aspects of life or the environment

No Comments

No Known Activations