INDEX

Explanations

statements expressing strong personal opinions

oai_token-act-pair · gpt-3.5-turbo

expressions of strong opinions 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

 Chaser

-0.86

adr

-0.76

 Journals

-0.74

ICAN

-0.74

 Victims

-0.71

 Gorge

-0.70

onut

-0.69

ILLE

-0.69

OTOS

-0.69

mits

-0.69

POSITIVE LOGITS

 strongly

0.99

 enough

0.94

 correlated

0.86

enough

0.82

 disagree

0.79

 unsu

0.76

 differentiated

0.76

 discrim

0.76

 distingu

0.74

 suggestive

0.73

Activations Density 0.004%

statements expressing strong personal opinions

expressions of strong opinions or sentiments

No Comments

No Known Activations