INDEX

Explanations

numerical data related to statistics and figures

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

irie

-0.74

imaru

-0.66

 sourcing

-0.65

uras

-0.63

hub

-0.61

worldly

-0.59

 medi

-0.59

 Galile

-0.58

 showc

-0.58

 contraction

-0.57

POSITIVE LOGITS

percent

0.83

0.80

0.76

degree

0.75

0.74

lvl

0.72

generic

0.70

Activations Density 0.094%

numerical data related to statistics and figures

No Comments

No Known Activations