INDEX

Explanations

sequences or remarks indicating negative assessments or declines in quality or performance

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

ppe

-0.82

conservancy

-0.80

ysis

-0.73

lene

-0.71

 proportion

-0.68

pport

-0.64

teenth

-0.63

bley

-0.62

 reper

-0.62

iod

-0.62

POSITIVE LOGITS

ADVERTISEMENT

0.82

––

0.81

 srfAttach

0.81

=-=-=-=-

0.78

Rated

0.75

=-=-

0.72

 Rapp

0.71

ĸļ

0.71

 Turns

0.70

å¥³

0.70

Activations Density 0.008%

sequences or remarks indicating negative assessments or declines in quality or performance

No Comments

No Known Activations