INDEX

Explanations

time-related phrases like "less than" and "just."

oai_token-act-pair · gpt-3.5-turbo

phrases indicating short time frames or quick actions

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

ourge

-0.70

apego

-0.67

 congratulated

-0.64

aily

-0.63

illard

-0.61

orem

-0.58

yss

-0.58

sers

-0.58

 underestimate

-0.57

 inhibitors

-0.57

POSITIVE LOGITS

 guise

0.91

 nutshell

0.85

 manner

0.82

 terms

0.79

 increments

0.79

 fashion

0.77

 regard

0.75

 context

0.75

 form

0.71

Shape

0.71

Activations Density 0.149%

time-related phrases like "less than" and "just."

phrases indicating short time frames or quick actions

No Comments

No Known Activations