INDEX

Explanations

references to specific states

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

schild

-0.89

sbm

-0.88

matter

-0.85

diagn

-0.74

ajor

-0.73

ãĤ¨ãĥ«

-0.72

ilater

-0.72

ily

-0.72

hran

-0.71

aird

-0.71

POSITIVE LOGITS

 Warriors

1.18

 Ferry

0.89

 Conference

0.83

 Sachs

0.77

 Pelicans

0.75

 Thunder

0.75

 Gauntlet

0.71

 Esports

0.71

 Lightning

0.70

 Lopez

0.69

Activations Density 0.011%

references to specific states

No Comments

No Known Activations