INDEX

Explanations

names of individuals, particularly those ending with "Patel" and "Cromartie."

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names related to individuals involved in specific events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

gression

-0.89

rants

-0.83

force

-0.80

ilater

-0.80

lisher

-0.78

alted

-0.77

amaz

-0.76

lator

-0.76

rights

-0.75

acement

-0.74

POSITIVE LOGITS

LTD

0.70

ength

0.69

¬¼

0.67

 Sabb

0.67

ORK

0.65

 nood

0.64

 Lich

0.63

 ease

0.61

 crank

0.61

 Galile

0.61

Activations Density 0.046%

names of individuals, particularly those ending with "Patel" and "Cromartie."

proper nouns, particularly names related to individuals involved in specific events

No Comments

No Known Activations