INDEX

Explanations

references to quantities, amounts, or notions of lack and presence

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

 fame

-0.67

 vouchers

-0.64

 Replacement

-0.58

 encouragement

-0.56

cov

-0.56

ioch

-0.56

 Owners

-0.56

atts

-0.56

 largeDownload

-0.55

veyard

-0.55

POSITIVE LOGITS

 aspects

0.83

 existing

0.78

 downstream

0.77

 facets

0.76

ones

0.75

heartedly

0.73

IRE

0.73

 havoc

0.72

 longstanding

0.72

orously

0.70

Activations Density 0.139%

references to quantities, amounts, or notions of lack and presence

No Comments

No Known Activations