INDEX

Explanations

instances of the word "ild."

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

andise

-0.76

SOURCE

-0.73

CHR

-0.69

okin

-0.66

 Dragonbound

-0.66

ãĤ®

-0.66

displayText

-0.64

meal

-0.64

éŃĶ

-0.63

ngth

-0.63

POSITIVE LOGITS

erers

0.94

reth

0.89

ered

0.82

rance

0.81

roid

0.80

sburg

0.80

ild

0.80

er

0.79

ings

0.78

Activations Density 0.004%

instances of the word "ild."

No Comments

No Known Activations