INDEX

Explanations

hyphenated phrases where the second word has a negative connotation

oai_token-act-pair · gpt-3.5-turbo

references to high-speed and high-profile subjects or 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

 Creator

-0.78

ulhu

-0.77

Äĩ

-0.71

 Guru

-0.69

 Dull

-0.69

ÄŁ

-0.68

Cah

-0.68

ĸļ

-0.67

ļéĨĴ

-0.66

Wiz

-0.66

POSITIVE LOGITS

level

1.26

sounding

1.15

powered

1.14

ranking

1.13

enough

1.07

grade

1.06

sized

1.06

definition

1.06

performing

1.04

quality

1.02

Activations Density 0.051%

hyphenated phrases where the second word has a negative connotation

references to high-speed and high-profile subjects or events

No Comments

No Known Activations