INDEX

Explanations

proper nouns or names

oai_token-act-pair · gpt-3.5-turbo

the character "K."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 understatement

-0.60

 Romanian

-0.57

 agric

-0.57

 stat

-0.56

 summ

-0.56

 stuffing

-0.56

 nort

-0.55

 overw

-0.55

 actionGroup

-0.55

 defic

-0.55

POSITIVE LOGITS

3.51

KS

2.32

2.19

KI

2.09

KK

2.06

KR

2.06

KA

1.95

KER

1.89

KT

1.89

1.88

Activations Density 0.027%

proper nouns or names

the character "K."

No Comments

No Known Activations