INDEX

Explanations

phrases related to ways or methods

oai_token-act-pair · gpt-3.5-turbo

the word "which" in various contexts

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

§

-0.75

 srfAttach

-0.70

Pak

-0.68

Ģ

-0.68

AGES

-0.67

Rog

-0.66

apt

-0.66

Cra

-0.66

rolet

-0.66

irt

-0.64

POSITIVE LOGITS

soever

0.87

xual

0.86

velt

0.73

 adaptation

0.70

 they

0.68

 humans

0.67

 organisms

0.67

rity

0.66

 liqu

0.65

eve

0.65

Activations Density 0.036%

phrases related to ways or methods

the word "which" in various contexts

No Comments

No Known Activations