INDEX

Explanations

the phrase "touched down in."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "wo" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

idates

-0.75

 Sunshine

-0.71

 mileage

-0.70

 Immunity

-0.69

 dissolve

-0.67

 Democracy

-0.67

death

-0.67

 Uriel

-0.66

ablishment

-0.65

matic

-0.64

POSITIVE LOGITS

wo

1.31

efully

1.22

ollen

1.12

eful

1.08

ocom

1.01

wo

0.97

jo

0.95

oks

0.86

olly

0.85

ofer

0.84

Activations Density 0.005%

the phrase "touched down in."

occurrences of the word "wo" in various contexts

No Comments

No Known Activations