INDEX

Explanations

capitalized place names

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "Ne" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

MODE

-0.86

displayText

-0.80

loo

-0.79

DOWN

-0.78

è¦ļéĨĴ

-0.77

Ranked

-0.76

ãĤ¼ãĤ¦ãĤ¹

-0.75

^^^^

-0.73

hips

-0.70

UID

-0.68

POSITIVE LOGITS

braska

1.08

arest

1.03

utral

1.00

Ne

0.98

cker

0.90

oliberal

0.85

XT

0.83

igen

0.80

olithic

0.79

hm

0.79

Activations Density 0.006%

capitalized place names

mentions of the word "Ne" in various contexts

No Comments

No Known Activations