INDEX

Explanations

instances where something is being clarified or explained

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "clarify" and its variants

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

aden

-0.80

©¶æ

-0.76

Leary

-0.72

heimer

-0.71

pes

-0.67

avery

-0.65

aea

-0.64

atri

-0.62

bay

-0.62

bearing

-0.62

POSITIVE LOGITS

 clar

1.30

 clarify

1.15

 clarified

1.13

 clarification

1.10

ifications

0.88

ifying

0.86

ifies

0.86

 deline

0.83

 EDIT

0.78

 misunderstand

0.77

Activations Density 0.008%

instances where something is being clarified or explained

occurrences of the word "clarify" and its variants

No Comments

No Known Activations