INDEX

Explanations

references to various rooms

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

words related to physical locations or environments

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 flakes

-0.69

 reconnect

-0.67

uras

-0.64

 Metall

-0.64

 Kickstarter

-0.61

 sacr

-0.61

 Americ

-0.61

 Ducks

-0.60

 Lois

-0.60

DPR

-0.59

POSITIVE LOGITS

room

0.98

idge

0.88

rooms

0.87

ãĤ£

0.86

tery

0.85

bial

0.81

bell

0.79

å§«

0.79

iday

0.78

sie

0.78

Activations Density 0.008%

references to various rooms

words related to physical locations or environments

No Comments

No Known Activations