INDEX

Explanations

verbs indicating actions or states of being in past and present tense

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

etheless

-0.81

äºĶ

-0.81

PDATE

-0.75

 disobedience

-0.69

ç·

-0.67

 somew

-0.66

theless

-0.65

issance

-0.65

å¼

-0.65

åĳ

-0.64

POSITIVE LOGITS

ilon

1.11

esley

0.90

iston

0.89

ams

0.89

ley

0.86

afe

0.82

endor

0.82

avan

0.82

aq

0.82

zel

0.80

Activations Density 0.070%

verbs indicating actions or states of being in past and present tense

No Comments

No Known Activations