INDEX

Explanations

names, specifically those related to places or organizations preceded by 'AL'

oai_token-act-pair · gpt-3.5-turbo

the presence of the word "AL" in varied contexts

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

 gent

-0.66

 dign

-0.61

 punk

-0.61

hirt

-0.60

 Schw

-0.59

 Widow

-0.59

 bushes

-0.59

 humor

-0.59

 puck

-0.59

 joking

-0.59

POSITIVE LOGITS

AL

3.97

ALS

2.71

ALE

2.03

ALLY

1.97

ALT

1.97

als

1.90

al

1.78

ally

1.67

ALL

1.66

ALD

1.60

Activations Density 0.013%

names, specifically those related to places or organizations preceded by 'AL'

the presence of the word "AL" in varied contexts

No Comments

No Known Activations