INDEX

Explanations

words related to critical evaluation or negativity

oai_token-act-pair · gpt-3.5-turbo

words related to blunders and mistakes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

HER

-0.83

Construct

-0.74

OLOGY

-0.71

geist

-0.70

UAL

-0.69

pour

-0.68

OTOS

-0.68

lectic

-0.68

EMENT

-0.67

RAL

-0.67

POSITIVE LOGITS

anches

1.08

ameless

1.07

anche

1.07

ossom

1.03

ogging

0.98

ocking

0.98

itting

0.97

ushes

0.97

itter

0.96

uffs

0.96

Activations Density 0.009%

words related to critical evaluation or negativity

words related to blunders and mistakes

No Comments

No Known Activations