INDEX

Explanations

phrases related to giving reasons or justifying actions

oai_token-act-pair · gpt-3.5-turbo

instances where the word "because" is used to introduce explanations or justifications

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

mint

-0.72

jet

-0.70

wn

-0.68

robe

-0.67

ns

-0.67

ax

-0.65

aid

-0.65

sey

-0.64

shaw

-0.64

str

-0.61

POSITIVE LOGITS

 nobody

0.91

 otherwise

0.91

hey

0.91

 there

0.88

 obviously

0.85

 they

0.83

 frankly

0.81

 fuck

0.81

we

0.79

it

0.77

Activations Density 0.086%

phrases related to giving reasons or justifying actions

instances where the word "because" is used to introduce explanations or justifications

No Comments

No Known Activations