INDEX

Explanations

mentions of a specific person, in this case, "Ford"

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

 Artemis

-0.75

legraph

-0.70

 certific

-0.69

ablishment

-0.67

 Alien

-0.67

anwhile

-0.66

 Sakuya

-0.65

 referen

-0.65

hemat

-0.63

 Seym

-0.62

POSITIVE LOGITS

ham

1.16

shire

0.97

ragon

0.96

bies

0.89

nec

0.79

isle

0.79

ring

0.78

zie

0.77

bie

0.76

clad

0.75

Activations Density 0.003%

mentions of a specific person, in this case, "Ford"

No Comments

No Known Activations