INDEX

Explanations

references to unidentified or unnamed individuals

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

arium

-0.79

ageddon

-0.79

akings

-0.77

opia

-0.76

byss

-0.76

efficient

-0.73

urgy

-0.73

lov

-0.72

alach

-0.71

blems

-0.69

POSITIVE LOGITS

 persons

1.04

 unnamed

1.03

 narrator

0.95

 sources

0.91

 individuals

0.90

 unidentified

0.89

 person

0.89

 assailants

0.89

jur

0.83

 member

0.82

Activations Density 0.011%

references to unidentified or unnamed individuals

No Comments

No Known Activations