INDEX

Explanations

conjunctions joining words or clauses

oai_token-act-pair · gpt-3.5-turbo

conjunctions and phrases that signal connections and relationships between ideas

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

gae

-0.68

Rain

-0.67

Fine

-0.65

 Protestant

-0.63

 Bots

-0.62

Boss

-0.61

 Sleeping

-0.61

 Domestic

-0.61

 Wond

-0.61

 Diesel

-0.60

POSITIVE LOGITS

 resultant

0.93

 ensuing

0.88

 subsequent

0.84

 corresponding

0.81

rium

0.78

 inaction

0.78

actions

0.77

 consequ

0.77

 accompanying

0.74

 eventual

0.73

Activations Density 0.294%

conjunctions joining words or clauses

conjunctions and phrases that signal connections and relationships between ideas

No Comments

No Known Activations