INDEX

Explanations

phrases describing initial actions, thoughts, or observations

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "first impressions" or initial reactions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

vind

-0.70

yet

-0.68

still

-0.67

aden

-0.66

recy

-0.63

RW

-0.63

GF

-0.61

lov

-0.61

claimer

-0.60

 Canad

-0.60

POSITIVE LOGITS

 responders

0.83

 noticed

0.82

 visitors

0.72

 encountered

0.71

asma

0.67

 inevitably

0.67

 introdu

0.67

 alerted

0.66

 noticing

0.66

 reaction

0.66

Activations Density 0.086%

phrases describing initial actions, thoughts, or observations

references to the concept of "first impressions" or initial reactions

No Comments

No Known Activations