INDEX

Explanations

words related to legal contexts or actions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 orcs

-0.58

 Whitman

-0.58

SourceFile

-0.57

 Borderlands

-0.56

 anus

-0.56

 Tycoon

-0.56

gey

-0.55

 stitch

-0.53

Austral

-0.53

brow

-0.53

POSITIVE LOGITS

 Lauder

1.29

lishes

0.78

vez

0.77

ateg

0.76

Ã©e

0.74

enza

0.74

Ã©s

0.74

ement

0.74

ADRA

0.72

ciating

0.69

Activations Density 0.014%

words related to legal contexts or actions

No Comments

No Known Activations