INDEX

Explanations

references to numerical values or figures

oai_token-act-pair · gpt-3.5-turbo

references to numerical data or statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ModLoader

-0.86

 Conditions

-0.72

Interstitial

-0.72

INTON

-0.72

Â¯Â¯

-0.69

 Shroud

-0.69

nee

-0.67

sole

-0.67

structed

-0.66

OPLE

-0.64

POSITIVE LOGITS

 numbers

0.88

 Numbers

0.82

pace

0.82

metry

0.82

paces

0.80

peak

0.74

hare

0.74

ynthesis

0.72

 averages

0.72

 digits

0.72

Activations Density 0.026%

references to numerical values or figures

references to numerical data or statistics

No Comments

No Known Activations