INDEX

Explanations

tech-related terms or references, specifically pertaining to messaging or communication apps

oai_token-act-pair · gpt-3.5-turbo

mentions of "WhatsApp"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

LECT

-0.74

æµ

-0.70

 Lomb

-0.69

SOURCE

-0.67

iling

-0.67

urally

-0.67

JM

-0.67

 Staten

-0.67

REF

-0.66

eb

-0.66

POSITIVE LOGITS

 Whats

1.07

 whats

0.91

soever

0.87

 happ

0.87

 lapt

0.86

creen

0.86

hower

0.84

 cannabin

0.83

ilon

0.80

 artif

0.80

Activations Density 0.008%

tech-related terms or references, specifically pertaining to messaging or communication apps

mentions of "WhatsApp"

No Comments

No Known Activations