INDEX

Explanations

words indicating a strong negative opinion or action

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

crow

-0.72

essed

-0.65

 Pressure

-0.64

eman

-0.64

nings

-0.63

ammy

-0.61

hma

-0.60

ILY

-0.59

isson

-0.58

imir

-0.58

POSITIVE LOGITS

tering

0.80

 ç¥ŀ

0.73

:(

0.71

 guiActiveUn

0.71

ascript

0.71

ainer

0.69

heartedly

0.66

ingu

0.65

 except

0.64

onne

0.63

Activations Density 0.024%

words indicating a strong negative opinion or action

No Comments

No Known Activations