INDEX

Explanations

specific locations, likely in the context of a news article or report

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Ont" and related mentions of Ontario

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

ãģĦ

-1.01

ãĥ¼ãĥĨ

-0.88

ãĥ¬

-0.77

[+

-0.77

ãĤ§

-0.77

ãģį

-0.74

BLE

-0.73

ãĤ¼ãĤ¦ãĤ¹

-0.72

é¾įåĸļå£«

-0.70

BOOK

-0.69

POSITIVE LOGITS

ario

1.07

ologies

1.02

rack

1.00

rue

0.99

omet

0.97

ological

0.95

ologically

0.94

arios

0.93

omach

0.92

ology

0.91

Activations Density 0.008%

specific locations, likely in the context of a news article or report

occurrences of the word "Ont" and related mentions of Ontario

No Comments

No Known Activations