INDEX

Explanations

expressions involving swearing and emphasis

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

NetMessage

-0.94

KY

-0.76

 Cosponsors

-0.74

anwhile

-0.68

CRE

-0.68

ynthesis

-0.67

KEN

-0.65

OHN

-0.64

IDS

-0.64

chn

-0.64

POSITIVE LOGITS

ibly

0.89

 damned

0.84

 darn

0.83

ation

0.83

 damn

0.82

near

0.81

orse

0.79

atio

0.78

selves

0.76

warm

0.74

Activations Density 0.024%

expressions involving swearing and emphasis

No Comments

No Known Activations