INDEX

Explanations

phrases related to leaving a location or situation

oai_token-act-pair · gpt-3.5-turbo

mentions of various locations, especially fields and countries

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

sbm

-0.70

Gaw

-0.69

Ô

-0.64

 reconc

-0.62

 plent

-0.61

tein

-0.61

iqueness

-0.60

 Ning

-0.59

haps

-0.58

ĻĤ

-0.58

POSITIVE LOGITS

osphere

0.80

forts

0.75

 courthouse

0.71

ifice

0.71

places

0.70

walk

0.70

 safely

0.68

waters

0.67

 floor

0.66

washer

0.66

Activations Density 0.239%

phrases related to leaving a location or situation

mentions of various locations, especially fields and countries

No Comments

No Known Activations