INDEX

Explanations

reflexive pronouns

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

osi

-0.77

iversal

-0.75

terness

-0.73

 Seym

-0.72

odies

-0.71

ibaba

-0.70

vez

-0.70

ccording

-0.68

ione

-0.63

yip

-0.63

POSITIVE LOGITS

åĳ

0.69

or

0.67

 alone

0.63

 istg

0.63

ford

0.62

 overnight

0.61

 bare

0.61

EngineDebug

0.60

iHUD

0.58

Activations Density 0.020%

reflexive pronouns

No Comments

No Known Activations