INDEX

Explanations

the presence of specific names and titles associated with roles or organizations

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

 Beir

-0.78

orb

-0.72

utical

-0.67

HAEL

-0.66

 Finn

-0.66

CHAT

-0.66

english

-0.63

hovah

-0.63

Hav

-0.61

 charism

-0.61

POSITIVE LOGITS

inct

0.90

ous

0.84

igent

0.80

gence

0.79

ils

0.78

cius

0.76

ts

0.76

gent

0.76

²¾

0.75

ations

0.74

Activations Density 0.005%

the presence of specific names and titles associated with roles or organizations

No Comments

No Known Activations