INDEX

Explanations

language related to political news and analysis

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

qus

-0.76

_.

-0.68

 latter

-0.67

beit

-0.64

 triv

-0.64

 theirs

-0.64

 endif

-0.61

$.

-0.60

«ĺ

-0.59

ãĥĥãĥī

-0.58

POSITIVE LOGITS

 regarding

0.91

 Canaver

0.82

hillary

0.76

 relating

0.74

 concerning

0.72

 Regarding

0.71

 pertaining

0.69

 legalizing

0.67

 Donald

0.67

UFC

0.65

Activations Density 1.089%

language related to political news and analysis

No Comments

No Known Activations