INDEX

Explanations

linear relationships or patterns

oai_token-act-pair · gpt-3.5-turbo

terms associated with linear and exponential mathematical concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

AMI

-0.74

Downloadha

-0.73

UNE

-0.73

egu

-0.70

aco

-0.69

illon

-0.68

Standing

-0.68

ardless

-0.68

ARDS

-0.67

IRED

-0.67

POSITIVE LOGITS

 algebra

0.97

 interpol

0.93

ization

0.90

ized

0.84

izes

0.82

izable

0.79

icity

0.79

 progression

0.79

 linear

0.78

gebra

0.77

Activations Density 0.010%

linear relationships or patterns

terms associated with linear and exponential mathematical concepts

No Comments

No Known Activations