INDEX

Explanations

expressions of surprise or disbelief

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

 segreg

-0.67

 elim

-0.66

 Lesbian

-0.63

 Personality

-0.62

 Luther

-0.61

 Feld

-0.61

 Liberia

-0.60

 Townsend

-0.60

Spa

-0.60

 Pixie

-0.58

POSITIVE LOGITS

esome

1.57

akening

1.28

kward

1.16

alls

1.03

ards

1.00

iring

0.95

ake

0.92

orks

0.91

aw

0.91

reck

0.90

Activations Density 0.009%

expressions of surprise or disbelief

No Comments

No Known Activations