INDEX

Explanations

phrases related to reports, studies, and evidence in a scientific context

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

venge

-0.66

 Mouth

-0.62

Interstitial

-0.61

$.

-0.60

ga

-0.60

die

-0.59

ibles

-0.58

-+

-0.57

nice

-0.57

+.

-0.56

POSITIVE LOGITS

 suggests

1.05

 indicates

1.01

 indicate

0.99

uggest

0.97

 showed

0.97

 suggest

0.94

 indicated

0.93

 proved

0.90

 however

0.88

 confirms

0.85

Activations Density 0.172%

phrases related to reports, studies, and evidence in a scientific context

No Comments

No Known Activations