INDEX

Explanations

expressions or actions denoting agreement, approval, or acknowledgment

oai_token-act-pair · gpt-3.5-turbo

references to the act of nodding 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

inventoryQuantity

-0.84

 tabl

-0.83

 Journals

-0.72

OTOS

-0.71

 Hebdo

-0.69

 Tycoon

-0.67

Cov

-0.66

ÄŁ

-0.65

ãĥł

-0.65

Ļ½

-0.65

POSITIVE LOGITS

ules

1.14

nod

1.10

endum

1.05

ule

1.04

atory

1.03

uling

0.94

ularity

0.92

ulously

0.91

emon

0.89

ename

0.88

Activations Density 0.015%

expressions or actions denoting agreement, approval, or acknowledgment

references to the act of nodding in various contexts

No Comments

No Known Activations