INDEX

Explanations

references to the name "Kin" with varying levels of specificity and relevance

oai_token-act-pair · gpt-3.5-turbo

references to the term "kin" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

theless

-0.94

enance

-0.93

 Seym

-0.84

shire

-0.79

Ö¼

-0.77

ORD

-0.67

ERE

-0.67

×Ļ

-0.66

enegger

-0.64

 Maduro

-0.64

POSITIVE LOGITS

folk

1.25

etics

1.17

etic

1.17

sey

1.15

Kin

1.11

esthetic

1.07

etically

1.06

zie

1.06

ners

0.95

sell

0.93

Activations Density 0.020%

references to the name "Kin" with varying levels of specificity and relevance

references to the term "kin" in various contexts

No Comments

No Known Activations