INDEX

Explanations

words related to interruptions or interjections in conversation, particularly starting with "Oh," and ", and"

oai_token-act-pair · gpt-3.5-turbo

repeated phrases or expressions of surprise or emphasis

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

ascript

-0.86

ļéĨĴ

-0.70

resso

-0.66

İĭ

-0.66

ome

-0.65

inction

-0.64

emo

-0.64

UGC

-0.63

robe

-0.63

ForgeModLoader

-0.63

POSITIVE LOGITS

 yeah

1.05

yes

0.93

sir

0.85

 thank

0.82

 sorry

0.82

uh

0.81

 dear

0.79

 Wait

0.77

 hello

0.75

 please

0.73

Activations Density 0.103%

words related to interruptions or interjections in conversation, particularly starting with "Oh," and ", and"

repeated phrases or expressions of surprise or emphasis

No Comments

No Known Activations