INDEX

Explanations

phrases related to political or social issues in specific contexts

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.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

 JPEG

-0.73

 Maced

-0.66

 simulated

-0.65

 counted

-0.63

 misunder

-0.63

 decimal

-0.62

 mathemat

-0.61

 fortun

-0.60

 Axel

-0.59

 decomp

-0.59

POSITIVE LOGITS

0.99

_.

0.92

ski

0.91

tion

0.90

stra

0.88

ï¸ı

0.84

span

0.83

sure

0.82

edi

0.81

him

0.80

Activations Density 0.249%

phrases related to political or social issues in specific contexts

No Comments

No Known Activations