INDEX

Explanations

words or phrases associated with areas or regions, often linked to names or identities

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

FORMATION

-0.79

FTWARE

-0.68

 refere

-0.62

<+

-0.58

 rapt

-0.58

 Jury

-0.56

prises

-0.55

osexual

-0.55

kan

-0.55

ocry

-0.55

POSITIVE LOGITS

dash

0.85

ieri

0.81

anova

0.74

ioxide

0.74

elli

0.73

istan

0.72

Ã£o

0.71

eros

0.71

aughs

0.69

za

0.68

Activations Density 0.022%

words or phrases associated with areas or regions, often linked to names or identities

No Comments

No Known Activations