INDEX

Explanations

instances of the word "one" with positive emphasis or related concepts

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "one" in relationships and interactions among individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĤ¤ãĥĪ

-0.76

ãĤ¬

-0.75

vier

-0.73

lov

-0.69

ãĥ©ãĥ³

-0.69

ãģĤ

-0.68

è¦ļéĨĴ

-0.68

grounds

-0.67

shed

-0.67

å§«

-0.65

POSITIVE LOGITS

 another

1.14

another

1.03

 sided

0.99

 side

0.90

 particular

0.90

's

0.89

 hundred

0.89

 dimensional

0.87

 thing

0.87

 hemisphere

0.85

Activations Density 0.080%

instances of the word "one" with positive emphasis or related concepts

references to the concept of "one" in relationships and interactions among individuals

No Comments

No Known Activations