INDEX

Explanations

proper nouns, such as names of locations and people

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically locations and their associated attributes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

lihood

-0.79

::::::::

-0.72

pire

-0.67

inav

-0.65

insk

-0.65

cean

-0.64

perse

-0.64

amia

-0.63

chwitz

-0.63

advertisement

-0.63

POSITIVE LOGITS

 Mayor

1.40

 residents

1.28

 officials

1.21

 lawmakers

1.13

 politicians

1.12

 voters

1.10

 authorities

1.10

 mayor

1.09

 resident

1.06

 mayors

1.06

Activations Density 0.266%

proper nouns, such as names of locations and people

proper nouns, specifically locations and their associated attributes

No Comments

No Known Activations