INDEX

Explanations

words related to exaggeration or simplification

oai_token-act-pair · gpt-3.5-turbo

instances of the word "overs" or its variations, indicating a focus on the concept of oversimplification

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

arc

-0.71

 Nanto

-0.68

 motions

-0.63

 corridor

-0.62

Sho

-0.62

Dot

-0.62

Guard

-0.62

 utmost

-0.61

 theater

-0.61

 tooth

-0.60

POSITIVE LOGITS

 overs

1.29

impl

1.14

aturated

0.98

lap

0.95

leep

0.90

amples

0.86

icro

0.84

laughter

0.80

ummer

0.78

olicited

0.76

Activations Density 0.005%

words related to exaggeration or simplification

instances of the word "overs" or its variations, indicating a focus on the concept of oversimplification

No Comments

No Known Activations