INDEX

Explanations

phrases indicating a connection or engagement with a concept or idea

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate engagement or investment in ideas or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

eller

-0.79

erenn

-0.78

anooga

-0.74

enegger

-0.72

cot

-0.70

lege

-0.68

ĸļ

-0.68

bender

-0.68

warning

-0.68

antry

-0.67

POSITIVE LOGITS

 infinity

0.79

 existence

0.69

 adulthood

0.68

 eternity

0.66

 twitch

0.65

 submission

0.62

ffee

0.60

 profitability

0.60

 knots

0.60

 hypers

0.59

Activations Density 0.067%

phrases indicating a connection or engagement with a concept or idea

phrases that indicate engagement or investment in ideas or situations

No Comments

No Known Activations