INDEX

Explanations

text descriptions and explanations, potentially related to graphs, scientific articles, and images

oai_token-act-pair · gpt-3.5-turbo

sections that provide statistical data or analysis

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 choking

-0.82

 cons

-0.81

 administr

-0.78

 choked

-0.75

 abst

-0.75

 apost

-0.74

 appoint

-0.71

 neut

-0.71

 administered

-0.70

 babys

-0.70

POSITIVE LOGITS

Screenshot

1.45

Notice

1.35

Figure

1.26

Courtesy

1.26

Above

1.25

Pict

1.21

Below

1.17

Conclusion

1.14

Looks

1.12

Video

1.08

Activations Density 0.422%

text descriptions and explanations, potentially related to graphs, scientific articles, and images

sections that provide statistical data or analysis

No Comments

No Known Activations