INDEX

Explanations

time-related phrases indicating a long duration or history

oai_token-act-pair · gpt-3.5-turbo

mentions of periods of time, specifically focused on the concept of "years."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

XY

-0.71

apo

-0.64

aleigh

-0.63

 Shack

-0.62

guyen

-0.62

Sov

-0.60

arty

-0.60

HQ

-0.60

emort

-0.60

agraph

-0.59

POSITIVE LOGITS

ago

1.09

 past

0.92

Ago

0.72

long

0.69

mith

0.68

manship

0.68

chool

0.66

 passed

0.66

ional

0.66

 thereafter

0.64

Activations Density 0.057%

time-related phrases indicating a long duration or history

mentions of periods of time, specifically focused on the concept of "years."

No Comments

No Known Activations