INDEX

Explanations

references to reports or articles that feature the phrase "in."

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

wcs

-0.79

ãĤ¼ãĤ¦ãĤ¹

-0.73

orea

-0.72

erate

-0.69

enez

-0.68

emate

-0.68

 lasted

-0.66

annon

-0.65

agos

-0.65

orean

-0.64

POSITIVE LOGITS

 subsection

0.73

 Hacker

0.72

 subsections

0.69

 Lonely

0.69

 paragraphs

0.68

 Suff

0.67

 Mish

0.67

 Slate

0.67

 Watts

0.66

[|

0.66

Activations Density 0.055%

references to reports or articles that feature the phrase "in."

No Comments

No Known Activations