INDEX

Explanations

references to experimental research or studies

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

veland

-0.80

die

-0.76

iris

-0.75

atra

-0.75

si

-0.74

WHERE

-0.73

olulu

-0.71

kins

-0.71

andra

-0.70

criptions

-0.70

POSITIVE LOGITS

imental

0.97

ists

0.87

ization

0.84

ized

0.79

 Prototype

0.77

izations

0.77

 explor

0.75

 Experimental

0.72

ally

0.72

izing

0.71

Activations Density 0.008%

references to experimental research or studies

No Comments

No Known Activations