INDEX

Explanations

phrases containing the expression "go-to"

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate essential actions or recommendations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Rivers

-0.72

 Petition

-0.65

Dial

-0.60

ĩ

-0.59

rots

-0.59

olls

-0.58

akespeare

-0.58

ollo

-0.57

ysics

-0.57

clerosis

-0.56

POSITIVE LOGITS

terday

0.89

ãĤ´ãĥ³

0.78

iste

0.77

gotten

0.75

 superpower

0.72

zzle

0.71

theless

0.71

ļéĨĴ

0.67

kered

0.66

glers

0.63

Activations Density 0.056%

phrases containing the expression "go-to"

phrases that indicate essential actions or recommendations

No Comments

No Known Activations