INDEX

Explanations

adverbs, particularly those ending in 'ly'

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

 GOODMAN

-0.92

ilater

-0.85

afety

-0.84

irlf

-0.83

eport

-0.75

ERY

-0.73

 Emin

-0.72

 Extrem

-0.72

arella

-0.70

itures

-0.70

POSITIVE LOGITS

adv

0.97

 priced

0.89

gged

0.86

 supported

0.82

comed

0.79

present

0.78

tics

0.77

 rewarded

0.77

rics

0.76

 appreciated

0.74

Activations Density 0.031%

adverbs, particularly those ending in 'ly'

No Comments

No Known Activations