INDEX

Explanations

phrases indicating statements or claims

oai_token-act-pair · gpt-3.5-turbo

phrases that include claims or assertions about events or states of being

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

intosh

-0.67

irez

-0.67

emort

-0.63

pite

-0.63

patch

-0.61

DOWN

-0.61

leground

-0.60

ortunate

-0.60

 PLUS

-0.59

 Reconstruction

-0.58

POSITIVE LOGITS

 behave

0.79

 embody

0.76

esty

0.72

 manipulate

0.71

 perform

0.71

asted

0.70

 satisfy

0.68

 adhere

0.67

ads

0.67

 speak

0.66

Activations Density 0.105%

phrases indicating statements or claims

phrases that include claims or assertions about events or states of being

No Comments

No Known Activations