INDEX

Explanations

references to battlefields or combat zones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Wan

-0.78

pert

-0.73

yu

-0.72

andro

-0.70

perm

-0.70

QU

-0.70

por

-0.70

gger

-0.69

ulous

-0.69

aga

-0.68

POSITIVE LOGITS

1.00

 battlefield

0.91

sburg

0.90

front

0.83

 commanders

0.80

 commander

0.77

roy

0.75

side

0.73

 combatants

0.72

halla

0.72

Activations Density 0.010%

references to battlefields or combat zones

No Comments

No Known Activations