INDEX

Explanations

the word "thorny" or words related to difficult and complex issues

oai_token-act-pair · gpt-3.5-turbo

terms related to obstacles or complications, particularly those that are recurring or problematic

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ession

-0.77

rador

-0.73

algia

-0.72

ysis

-0.71

ONT

-0.71

olid

-0.70

rontal

-0.69

regation

-0.67

estation

-0.67

glomer

-0.66

POSITIVE LOGITS

 thorn

0.96

combe

0.85

baum

0.79

 Ashe

0.78

0.74

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.70

\\\\\\\\\\\\\\\\

0.68

cliffe

0.68

eas

0.68

Ô

0.67

Activations Density 0.023%

the word "thorny" or words related to difficult and complex issues

terms related to obstacles or complications, particularly those that are recurring or problematic

No Comments

No Known Activations