INDEX

Explanations

proper names followed by colons

oai_token-act-pair · gpt-3.5-turbo

statements and responses in a conversational or question-and-answer format

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 downstream

-0.69

 doub

-0.67

 wrath

-0.64

 forgotten

-0.63

 unchecked

-0.63

abad

-0.63

 rule

-0.62

foreseen

-0.62

 trespass

-0.61

 attention

-0.61

POSITIVE LOGITS

 Exactly

1.15

 Yeah

1.07

Absolutely

0.98

 Absolutely

0.95

 Originally

0.94

Yes

0.91

Yeah

0.90

 Provided

0.88

 Firstly

0.86

Hmm

0.86

Activations Density 0.030%

proper names followed by colons

statements and responses in a conversational or question-and-answer format

No Comments

No Known Activations