INDEX

Explanations

references to letters or documents related to support or intentions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nels

-0.85

é¾

-0.76

»Ĵ

-0.67

staking

-0.66

arios

-0.64

 ILCS

-0.64

aky

-0.62

ãĥ¯ãĥ³

-0.62

odied

-0.61

pes

-0.61

POSITIVE LOGITS

 letter

0.79

mbuds

0.76

 Agreement

0.74

 petitions

0.73

 resignation

0.73

 Letter

0.73

 Letters

0.72

 Petition

0.72

Letter

0.72

letters

0.72

Activations Density 0.036%

references to letters or documents related to support or intentions

No Comments

No Known Activations