INDEX

Explanations

words related to implications or suggestions

oai_token-act-pair · gpt-3.5-turbo

words related to implications or suggesting conclusions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

eret

-0.76

"},{"

-0.74

foot

-0.68

 surfing

-0.67

ryu

-0.67

vre

-0.66

igo

-0.65

erate

-0.64

SEA

-0.64

meter

-0.63

POSITIVE LOGITS

 impl

3.96

impl

2.22

 Impl

1.60

Impl

1.32

 collapse

1.16

 unravel

1.07

 collapsing

1.05

expl

1.05

 expl

1.04

 collapses

0.95

Activations Density 0.019%

words related to implications or suggestions

words related to implications or suggesting conclusions

No Comments

No Known Activations