INDEX

Explanations

names of places, potentially university-related

oai_token-act-pair · gpt-3.5-turbo

words with the suffix 'id' or variations of 'id'

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

cla

-0.72

 merc

-0.68

ACTED

-0.65

 acknowled

-0.65

theless

-0.63

hospital

-0.62

 canal

-0.62

hower

-0.62

 weakest

-0.61

 faint

-0.59

POSITIVE LOGITS

wagen

0.82

 Jagu

0.72

nit

0.71

addle

0.71

kees

0.71

otos

0.69

inski

0.68

itzer

0.68

idon

0.68

estyle

0.67

Activations Density 0.085%

names of places, potentially university-related

words with the suffix 'id' or variations of 'id'

No Comments

No Known Activations