INDEX

Explanations

instances where someone or something responds to a situation or question

oai_token-act-pair · gpt-3.5-turbo

instances of the word "respond."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

cutting

-0.89

rome

-0.72

scar

-0.72

Bav

-0.70

marks

-0.68

metal

-0.67

 Thieves

-0.66

fi

-0.64

 Oblivion

-0.64

fall

-0.64

POSITIVE LOGITS

 responses

1.00

 respond

0.99

reply

0.92

 response

0.86

 responders

0.84

isson

0.83

ively

0.83

 Response

0.82

 thereto

0.81

 adequately

0.80

Activations Density 0.022%

instances where someone or something responds to a situation or question

instances of the word "respond."

No Comments

No Known Activations