INDEX

Explanations

mentions of the word "guilty" or its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

hower

-0.87

baugh

-0.68

croft

-0.67

 Sacrament

-0.67

opsis

-0.65

 Associates

-0.62

 departures

-0.62

 Countdown

-0.61

quartered

-0.60

POSITIVE LOGITS

inea

1.30

ppy

1.06

vernment

1.04

cci

1.04

ilt

0.97

arding

0.95

ilts

0.94

ile

0.92

asant

0.91

illas

0.90

Activations Density 0.017%

mentions of the word "guilty" or its variations in different contexts

No Comments

No Known Activations