INDEX

Explanations

nouns related to physical actions or scenes

oai_token-act-pair · gpt-3.5-turbo 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

Services

-0.82

Newsletter

-0.80

URA

-0.78

Pwr

-0.78

SEA

-0.76

risk

-0.73

ickr

-0.72

KR

-0.72

UA

-0.71

Sub

-0.70

POSITIVE LOGITS

ingly

0.86

ings

0.85

ers

0.79

eting

0.78

ed

0.77

ing

0.76

edly

0.75

ership

0.73

wards

0.70

quartered

0.69

Activations Density 0.232%

nouns related to physical actions or scenes

No Comments

No Known Activations