INDEX

Explanations

phrases indicating the end or conclusion

oai_token-act-pair · gpt-3.5-turbo

references to summarizing or concluding statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

cause

-0.72

Ranked

-0.70

onent

-0.66

sure

-0.61

éĥ

-0.60

gypt

-0.60

Demand

-0.60

":""},{"

-0.60

Cause

-0.60

 href

-0.60

POSITIVE LOGITS

 summarize

0.86

 wraps

0.81

 nutshell

0.80

 recap

0.76

 summarizes

0.76

 summar

0.76

 summarized

0.71

 completes

0.69

 prest

0.68

 Summary

0.67

Activations Density 0.196%

phrases indicating the end or conclusion

references to summarizing or concluding statements

No Comments

No Known Activations