INDEX

Explanations

years, specifically targeting the year 1993

oai_token-act-pair · gpt-3.5-turbo

mentions of the year 1993

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Tu

-0.70

semb

-0.69

hed

-0.68

lying

-0.65

tro

-0.62

Stream

-0.60

gery

-0.60

bors

-0.59

Adv

-0.58

 lightsaber

-0.58

POSITIVE LOGITS

ĸļ

0.93

å¹

0.85

-'

0.80

theless

0.67

 Rwanda

0.65

uncture

0.65

ãĤ¦ãĤ¹

0.65

 Leban

0.63

ãĥ£

0.63

 Created

0.62

Activations Density 0.021%

years, specifically targeting the year 1993

mentions of the year 1993

No Comments

No Known Activations