INDEX

Explanations

mentions of individuals in executive roles

oai_token-act-pair · gpt-3.5-turbo

references to executive roles or positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Reincarn

-0.78

ologies

-0.75

lust

-0.74

Ô

-0.72

INGTON

-0.71

kn

-0.69

 Parables

-0.68

atoon

-0.68

some

-0.67

bane

-0.67

POSITIVE LOGITS

 director

1.18

 producer

1.18

 chef

1.04

 committee

1.04

 chairman

1.01

 vice

0.98

director

0.97

 assistant

0.94

 Producer

0.93

 editor

0.93

Activations Density 0.020%

mentions of individuals in executive roles

references to executive roles or positions

No Comments

No Known Activations