INDEX

Explanations

references to location or community settings

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

NECT

-0.72

pread

-0.72

æŃ

-0.69

zbek

-0.68

profile

-0.59

 Marketable

-0.58

ãĥŃ

-0.58

orsi

-0.58

Panel

-0.58

åĨ

-0.58

POSITIVE LOGITS

osphere

0.93

 itself

0.84

forts

0.78

 gods

0.73

elight

0.73

iest

0.71

 beforehand

0.70

ultimate

0.70

 tomorrow

0.69

 elves

0.68

Activations Density 0.244%

references to location or community settings

No Comments

No Known Activations