INDEX

Explanations

comparative adjectives

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

EVA

-0.77

ainted

-0.76

EP

-0.71

odor

-0.69

shire

-0.68

vp

-0.65

upuncture

-0.65

hr

-0.64

ulations

-0.64

syn

-0.63

POSITIVE LOGITS

 than

1.85

than

1.58

 Than

1.55

 versions

0.89

iating

0.88

 "$:/

0.81

 sibling

0.77

ado

0.75

 generations

0.73

 installments

0.72

Activations Density 0.548%

comparative adjectives

No Comments

No Known Activations