INDEX

Explanations

expressions related to extremes or intensity, such as "full savage" or "full steam"

oai_token-act-pair · gpt-3.5-turbo

intensity in actions or descriptions

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

Eg

-0.72

perture

-0.68

UCT

-0.66

Introduced

-0.64

ection

-0.64

 Wake

-0.62

Synopsis

-0.62

CRIPTION

-0.61

elligence

-0.60

 Valid

-0.60

POSITIVE LOGITS

 unnoticed

0.94

ĸļ

0.81

 downhill

0.78

Ô

0.77

ither

0.77

 smoothly

0.76

noticed

0.72

 undet

0.70

eper

0.69

 shopping

0.68

Activations Density 0.126%

expressions related to extremes or intensity, such as "full savage" or "full steam"

intensity in actions or descriptions

No Comments

No Known Activations