INDEX

Explanations

references to physical locations, particularly rooms, and related actions or attributes associated with those rooms

oai_token-act-pair · gpt-3.5-turbo

references to specific room numbers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

HER

-0.70

Iss

-0.69

elson

-0.67

pend

-0.66

merce

-0.65

Uz

-0.65

ernels

-0.58

gpu

-0.57

BAT

-0.57

 distribut

-0.56

POSITIVE LOGITS

room

1.02

 doors

0.92

 upstairs

0.91

rooms

0.90

 occupancy

0.89

 divid

0.89

itory

0.86

lain

0.84

 rooms

0.84

clock

0.83

Activations Density 0.060%

references to physical locations, particularly rooms, and related actions or attributes associated with those rooms

references to specific room numbers

No Comments

No Known Activations