INDEX

Explanations

positive or optimistic language

oai_token-act-pair · gpt-3.5-turbo

expressions emphasizing positive outcomes or sentiments

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

RAW

-0.92

arij

-0.79

ANS

-0.77

ITNESS

-0.74

jack

-0.72

oths

-0.72

ARS

-0.71

library

-0.70

Corn

-0.69

conservancy

-0.69

POSITIVE LOGITS

 positive

1.17

 reinforcement

1.01

positive

1.00

 Positive

0.98

 affirm

0.91

 affirmation

0.90

 feedback

0.88

negative

0.85

 Negative

0.85

 positives

0.85

Activations Density 0.015%

positive or optimistic language

expressions emphasizing positive outcomes or sentiments

No Comments

No Known Activations