INDEX

Explanations

references to second place or ranking

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "second."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

LOD

-0.72

aceae

-0.71

embed

-0.64

nuts

-0.63

avid

-0.62

HUD

-0.61

ains

-0.60

xual

-0.59

Features

-0.58

quit

-0.57

POSITIVE LOGITS

 second

3.29

second

2.62

 third

2.40

 fourth

2.35

Second

2.22

 Second

2.11

 fifth

2.07

 sixth

2.04

 seventh

1.97

 secondly

1.93

Activations Density 0.030%

references to second place or ranking

occurrences of the word "second."

No Comments

No Known Activations