INDEX

Explanations

references to sandwich-related terms

oai_token-act-pair · gpt-3.5-turbo

references to sandwiches

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

tics

-0.79

mberg

-0.75

ibr

-0.71

aft

-0.69

orne

-0.69

negative

-0.66

effic

-0.66

pmwiki

-0.66

axis

-0.65

umen

-0.63

POSITIVE LOGITS

 sandwiches

1.15

 sandwich

1.04

wich

1.01

bowl

0.91

 salads

0.91

 Sandwich

0.87

 tray

0.80

anut

0.80

 slices

0.80

 salad

0.79

Activations Density 0.012%

references to sandwich-related terms

references to sandwiches

No Comments

No Known Activations