INDEX

Explanations

instances of the word "goals."

oai_token-act-pair · gpt-3.5-turbo

references to objectives or targets 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oute

-0.72

AE

-0.67

Americ

-0.67

amination

-0.67

alter

-0.65

 Virus

-0.65

Doctors

-0.65

 Geological

-0.64

Otherwise

-0.64

 Hath

-0.63

POSITIVE LOGITS

 goals

1.08

 goal

0.97

 Goal

0.94

 Goals

0.94

Goal

0.89

posts

0.88

goal

0.87

 scorer

0.86

 kicker

0.84

hooting

0.80

Activations Density 0.014%

instances of the word "goals."

references to objectives or targets in various contexts

No Comments

No Known Activations