INDEX

Explanations

phrases indicating relative positioning or comparisons

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

 Cosponsors

-0.71

à¥

-0.71

BILL

-0.70

SHARE

-0.70

externalActionCode

-0.69

 istg

-0.68

drivers

-0.66

FIG

-0.66

du

-0.66

cu

-0.65

POSITIVE LOGITS

 darkness

0.87

 nowhere

0.86

 civilization

0.86

 iceberg

0.84

 civilisation

0.83

 cliff

0.82

 exhaustion

0.82

 plateau

0.81

 obscurity

0.81

 cliffs

0.79

Activations Density 0.037%

phrases indicating relative positioning or comparisons

No Comments

No Known Activations