INDEX

Explanations

adjectives followed by a hyphen and a number indicating degree

oai_token-act-pair · gpt-3.5-turbo

phrases related to negative conditions or criticisms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

lisher

-0.70

 Authors

-0.65

 antioxid

-0.65

ħĭ

-0.64

£ı

-0.64

 guiActiveUnfocused

-0.64

illon

-0.63

 scissors

-0.62

 Walls

-0.61

 "$:/

-0.60

POSITIVE LOGITS

gotten

0.96

iquid

0.77

usive

0.76

untled

0.71

imm

0.67

utton

0.67

inen

0.66

ventures

0.65

nered

0.63

earth

0.63

Activations Density 0.044%

adjectives followed by a hyphen and a number indicating degree

phrases related to negative conditions or criticisms

No Comments

No Known Activations