INDEX

Explanations

words related to communication and interaction between individuals or entities

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

iru

-0.67

oiler

-0.65

 invincible

-0.64

throw

-0.62

iris

-0.61

 Citation

-0.60

hog

-0.60

 Shutterstock

-0.60

 invaders

-0.59

ption

-0.59

POSITIVE LOGITS

 privately

1.06

 internally

1.05

 with

0.94

 beforehand

0.92

 inform

0.91

between

0.91

 regarding

0.90

 briefed

0.86

nesday

0.86

 between

0.86

Activations Density 0.277%

words related to communication and interaction between individuals or entities

No Comments

No Known Activations