INDEX

Explanations

incremental steps in numerical values, possibly related to measurements or ratings

oai_token-act-pair · gpt-3.5-turbo

instances of the word "us" 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 @ 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

Negative Logits

RAG

-0.71

td

-0.64

ottest

-0.61

ITNESS

-0.60

rought

-0.57

 reservations

-0.56

ested

-0.55

irst

-0.55

owler

-0.54

 payday

-0.54

POSITIVE LOGITS

pex

1.24

CRIP

1.15

chwitz

1.14

cular

1.09

peed

1.01

cules

1.01

sein

1.01

etts

0.99

henko

0.97

cul

0.96

Activations Density 0.076%

incremental steps in numerical values, possibly related to measurements or ratings

instances of the word "us" in various contexts

No Known Activations