INDEX

Explanations

references to political figures and events

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to media outlets and sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

 haun

-0.72

 bonding

-0.64

 hiber

-0.63

 awake

-0.63

 debates

-0.62

 bond

-0.62

 disturbed

-0.61

 demol

-0.61

 fasc

-0.60

culus

-0.60

POSITIVE LOGITS

Getty

1.62

GET

1.36

File

1.14

Files

1.03

 Getty

0.97

License

0.96

KEN

0.90

Images

0.90

Bloomberg

0.88

Reuters

0.88

Activations Density 0.027%

references to political figures and events

references to media outlets and sources

No Comments

No Known Activations