INDEX

Explanations

words related to objectives or targets to achieve

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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

amination

-0.74

fecture

-0.72

 largeDownload

-0.67

uge

-0.66

 Zimmer

-0.65

fect

-0.64

show

-0.64

 Colbert

-0.63

ochet

-0.63

instance

-0.63

POSITIVE LOGITS

 goals

3.60

 Goals

2.85

 goal

2.52

 objectives

2.32

goal

2.31

 Goal

2.10

Goal

2.09

 agendas

1.67

 aims

1.67

 ambitions

1.63

Activations Density 0.018%

words related to objectives or targets to achieve

references to objectives or targets in various contexts

No Comments

No Known Activations