INDEX

Explanations

acronyms with the letters "LE" in them, often with a numerical value connected to each

oai_token-act-pair · gpt-3.5-turbo

references to the term "LE" in a variety of contexts

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

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

ãĥ¼ãĥĨãĤ£

-0.83

ãĤ¢ãĥ«

-0.83

ãĤ±

-0.79

raints

-0.79

orial

-0.78

webkit

-0.76

ilities

-0.75

ouls

-0.75

iaries

-0.75

raint

-0.74

POSITIVE LOGITS

VEL

1.09

IGH

1.07

LE

1.02

ASE

0.90

WOOD

0.88

LOAD

0.88

ACH

0.87

BO

0.87

lean

0.86

MY

0.85

Activations Density 0.008%

acronyms with the letters "LE" in them, often with a numerical value connected to each

references to the term "LE" in a variety of contexts

No Comments

No Known Activations