INDEX

Explanations

keywords referring to proximity or examination in detail

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

ICAN

-0.86

AIN

-0.74

NZ

-0.71

heim

-0.66

Sov

-0.65

 Freak

-0.65

urers

-0.65

uria

-0.64

oat

-0.64

ains

-0.64

POSITIVE LOGITS

 proximity

1.17

 resemblance

0.97

sighted

0.87

 minded

0.85

 relatives

0.84

 ties

0.84

 scrutiny

0.83

 enough

0.82

 confid

0.81

 shave

0.81

Activations Density 0.378%

keywords referring to proximity or examination in detail

No Comments

No Known Activations