INDEX

Explanations

phrases indicating responses or reactions, particularly those involving dialogues or replies

oai_token-act-pair · gpt-3.5-turbo

instances of the word "with" in relation to actions or events

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

 Unloaded

-0.71

burst

-0.67

itute

-0.67

there

-0.64

usa

-0.64

taker

-0.61

wake

-0.61

chart

-0.59

main

-0.59

afia

-0.58

POSITIVE LOGITS

 impunity

1.14

 regards

1.08

 gust

1.04

vig

0.98

 regard

0.97

 respect

0.92

 caution

0.86

 slogans

0.85

 tales

0.83

 sarc

0.83

Activations Density 0.134%

phrases indicating responses or reactions, particularly those involving dialogues or replies

instances of the word "with" in relation to actions or events

No Comments

No Known Activations