INDEX

Explanations

phrases or sentences that introduce a contrast or counterpoint

oai_token-act-pair · gpt-3.5-turbo

the word "that" used in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

å§«

-0.75

sbm

-0.74

byss

-0.72

pots

-0.72

pill

-0.72

pecially

-0.72

hops

-0.71

 dstg

-0.71

olics

-0.71

adelphia

-0.71

POSITIVE LOGITS

 doesn

1.00

 ignores

0.93

 hasn

0.92

 doesnt

0.90

 nonetheless

0.89

 depends

0.89

 differs

0.86

isn

0.85

 didn

0.84

 wasn

0.84

Activations Density 0.106%

phrases or sentences that introduce a contrast or counterpoint

the word "that" used in various contexts

No Comments

No Known Activations