INDEX

Explanations

phrases related to logical thinking and problem-solving

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

Merit

-0.63

iations

-0.61

 Peninsula

-0.61

 camer

-0.60

çĦ

-0.60

 filming

-0.59

iona

-0.58

ç«

-0.58

ances

-0.58

feeding

-0.58

POSITIVE LOGITS

 about

0.93

 ABOUT

0.92

 aloud

0.83

about

0.79

 provoking

0.77

-|

0.76

estyle

0.74

ileaks

0.74

ãĤ¦ãĤ¹

0.73

 critically

0.73

Activations Density 0.091%

phrases related to logical thinking and problem-solving

No Comments

No Known Activations