INDEX

Explanations

the word "purely"

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a lack of impartiality or objectivity

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Mushroom

-0.78

lain

-0.78

è¦ļéĨĴ

-0.77

 Cards

-0.75

 Seeds

-0.75

 Neighbor

-0.72

 Nurs

-0.70

 Peninsula

-0.68

Doctors

-0.67

 Colleges

-0.67

POSITIVE LOGITS

 purely

0.95

reet

0.83

odox

0.82

pure

0.76

 minded

0.76

 speculative

0.76

technical

0.75

 guiActiveUn

0.74

rative

0.73

=#

0.72

Activations Density 0.003%

the word "purely"

phrases indicating a lack of impartiality or objectivity

No Comments

No Known Activations