INDEX

Explanations

adjectives related to appearance or quality

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

APD

-0.80

arta

-0.78

iment

-0.77

HAEL

-0.77

venant

-0.76

IAS

-0.72

iments

-0.71

upon

-0.71

KY

-0.70

JV

-0.69

POSITIVE LOGITS

 darn

1.27

 nifty

1.03

 neat

0.95

 much

0.94

 damn

0.90

 harmless

0.89

 snipp

0.89

 straightforward

0.88

 cool

0.88

 innocuous

0.85

Activations Density 0.265%

adjectives related to appearance or quality

No Comments

No Known Activations