INDEX

Explanations

words related to political or governmental themes

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

DATA

-0.71

llan

-0.65

Factor

-0.59

 Debor

-0.59

 Citizen

-0.59

natureconservancy

-0.58

 FANTASY

-0.56

ãĥ¼ãĥĨãĤ£

-0.56

Glass

-0.55

âĸĳâĸĳ

-0.55

POSITIVE LOGITS

ipop

1.19

ateral

1.14

phe

1.09

okia

1.06

onial

1.02

oad

0.95

ounge

0.95

ibrary

0.94

ocation

0.93

anguages

0.93

Activations Density 0.011%

words related to political or governmental themes

No Comments

No Known Activations