INDEX

Explanations

phrases related to media appearances or political statements

oai_token-act-pair · gpt-3.5-turbo

reported statements or quotes from interviews and shows

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 packing

-0.75

etsy

-0.72

alties

-0.67

 scrib

-0.66

 bloom

-0.66

 envelop

-0.65

 shipping

-0.65

 curs

-0.65

idas

-0.64

acca

-0.64

POSITIVE LOGITS

 interviewer

1.15

 Tonight

1.13

 radio

1.10

 aired

1.07

 MSNBC

1.06

 Hannity

1.05

 broadcaster

1.04

 CNBC

1.04

Radio

1.04

radio

1.02

Activations Density 0.292%

phrases related to media appearances or political statements

reported statements or quotes from interviews and shows

No Comments

No Known Activations