INDEX

Explanations

references to American football teams, especially the Atlanta Falcons

oai_token-act-pair · gpt-3.5-turbo

mentions of the Atlanta Falcons football team

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

cules

-0.79

 Kardashian

-0.77

bell

-0.73

pert

-0.72

zanne

-0.71

fare

-0.69

tics

-0.66

ocard

-0.66

gging

-0.64

fu

-0.63

POSITIVE LOGITS

 Falcons

1.14

 Braves

0.80

layer

0.77

ï¸

0.77

daq

0.71

 Buccaneers

0.69

BSD

0.69

manship

0.69

rette

0.68

 Hawks

0.66

Activations Density 0.019%

references to American football teams, especially the Atlanta Falcons

mentions of the Atlanta Falcons football team

No Comments

No Known Activations