INDEX

Explanations

phrases related to corruption and irregularities in different contexts

oai_token-act-pair · gpt-3.5-turbo

references to corruption and irregularities 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 @ 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

AM

-0.73

amy

-0.72

linger

-0.71

los

-0.69

acles

-0.67

am

-0.67

chall

-0.67

leans

-0.67

 knocks

-0.66

 sidx

-0.66

POSITIVE LOGITS

effic

0.91

efficiency

0.84

 situ

0.80

 Northeast

0.73

 lineback

0.72

 infancy

0.72

roads

0.70

appropri

0.68

 workplace

0.68

 regards

0.67

Activations Density 0.347%

phrases related to corruption and irregularities in different contexts

references to corruption and irregularities in various contexts

No Comments

No Known Activations