INDEX

Explanations

phrases related to contrasting ideas

oai_token-act-pair · gpt-3.5-turbo

statements questioning common beliefs or assumptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 advis

-0.77

 shake

-0.61

 unknown

-0.61

 washed

-0.60

 inspector

-0.58

apr

-0.58

 culmin

-0.58

 indec

-0.58

 shepherd

-0.58

 uncertain

-0.58

POSITIVE LOGITS

 Rather

1.59

Nor

1.50

 Instead

1.49

Rather

1.26

 Nope

1.22

Instead

1.21

nor

1.21

 Quite

1.18

Nor

1.12

nor

1.07

Activations Density 0.401%

phrases related to contrasting ideas

statements questioning common beliefs or assumptions

No Comments

No Known Activations