INDEX

Explanations

No Explanations Found

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

²¾

-0.75

 Faction

-0.70

 Residents

-0.69

 Cheong

-0.68

 intervene

-0.68

 wors

-0.67

 evict

-0.66

 ants

-0.64

oths

-0.63

ulhu

-0.63

POSITIVE LOGITS

ISC

0.76

andel

0.70

\\\\\\\\

0.69

pee

0.67

joy

0.67

anmar

0.64

iscovery

0.64

senal

0.64

apon

0.63

å½

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.