INDEX

Explanations

phrases indicating affiliation or connection to a particular team or group

oai_token-act-pair · gpt-3.5-turbo

clauses that include the word "that," indicating a focus on explanatory or descriptive phrases

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 Pets

-0.71

 Pict

-0.70

 Gallery

-0.67

 Dome

-0.66

 Cheong

-0.66

bats

-0.65

 Pavilion

-0.64

IVERS

-0.62

 Another

-0.62

 Directions

-0.61

POSITIVE LOGITS

lished

0.75

ozy

0.73

yssey

0.72

 plagued

0.71

 comprised

0.71

¿½

0.70

alore

0.69

 consumes

0.67

href

0.67

 convol

0.67

Activations Density 0.287%

phrases indicating affiliation or connection to a particular team or group

clauses that include the word "that," indicating a focus on explanatory or descriptive phrases

No Known Activations