INDEX

Explanations

specific references to published articles or news reports

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word 'the' and its variants within various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 thereof

-0.80

SPONSORED

-0.74

.''

-0.70

).

-0.68

!).

-0.66

.</

-0.66

!.

-0.66

 thereby

-0.65

.ï¿½

-0.65

POSITIVE LOGITS

 same

1.12

 simplest

1.09

 smallest

1.06

 latest

1.01

 entire

0.99

 largest

0.99

oret

0.98

 slightest

0.98

 entirety

0.97

 hardest

0.97

Activations Density 1.815%

specific references to published articles or news reports

occurrences of the word 'the' and its variants within various contexts

No Comments

No Known Activations