INDEX

Explanations

information related to individuals' backgrounds, careers, and accomplishments

oai_token-act-pair · gpt-3.5-turbo

connections between entities and their roles or achievements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ptoms

-0.73

>[

-0.71

rium

-0.67

rid

-0.67

odo

-0.67

actionDate

-0.67

yip

-0.66

Definition

-0.65

isse

-0.65

IRE

-0.64

POSITIVE LOGITS

 oversaw

1.08

 participated

1.06

 teaches

1.02

 toured

0.96

 specializes

0.95

 participates

0.93

 owns

0.91

 enjoys

0.89

 volunteered

0.88

 frequ

0.86

Activations Density 0.347%

information related to individuals' backgrounds, careers, and accomplishments

connections between entities and their roles or achievements

No Comments

No Known Activations