INDEX

Explanations

details related to legal and political controversies

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

igr

-0.68

oj

-0.68

":-

-0.67

EMENT

-0.67

oused

-0.66

igator

-0.66

uesday

-0.65

erd

-0.65

gat

-0.65

esan

-0.65

POSITIVE LOGITS

 possibly

0.92

 worse

0.87

cru

0.86

 downright

0.86

 incidentally

0.83

 vice

0.82

 guessed

0.81

 importantly

0.78

quite

0.77

 rightfully

0.77

Activations Density 2.752%

details related to legal and political controversies

No Comments

No Known Activations